The Chemical Production of Mutations

C. Auerbach, J. M. Robson, and J. G. Carr

Institate of Animal Genetics and Department of Pharmacology, University of Edinbargh

A
GENETIC MUTATION IS A CHANGE, mutpion. If, as we assume, a mutation is a chemical
presumably chemical in nature, in one of the process, then knowledge of the reagents capable of
genes which compose the chromosome thread. initiating this process should throw light not only on the
The mutated gene is as stable as the original; it goes on reaction itself, but also on the nature of the gene, the
reproducing replicas of its mutated self and thus initiates other partner in the reaction. Moreover, it could be
a new hereditary line. I t is believed that without hoped that among chemical mutagens there might be
mutation life would never have proceeded beyond its very some with particular affinities for individual genes.
first elementary beginnings. Net the mechanism of Detection of such substances not only would be of high
this important process is practically unknown. With theoretical interest but would also open up the long-
the discovery by H. J. Muller (12) that mutations can be sought-for way to the production of directed mutations.
produced artifically by X-rays, a new approach to the The search for chemical mutagens has been going on for
problem of mutation was opened up, and many new facts well over 20 years. The choice of substances tried for the
relating to this problem were brought to light. One of purpose was mainly random. Iodine, ammonia, metaI
the most important results was the discovery that the compounds, and carcinogens are only some out of the
term "mutation" includes a number of distinct processes. great number tested. Results were often clearly negative,
I n addition to gene mutations as defined above, X-rays and no clearly positive ones had been obtained up to the
produce breaks in the chromosome threads; when the beginning of our experiments in 1940 (see 1, 9, 19). It
resulting fragments join together in novel combinations, is obvious that a chemical mutagen must possess very
so-called chromosome rearrangements, hereditary special properties. I t must be able to act selectively on
changes closely resembling mutations may be produced. the genic material without at the same 'time destroying
Rules connecting quantity and quality of radiation with the cell which contains this material. I t was, therefore,
type and frequency of mutations were discovered, and only to be expected that many substances would have to
inferences could be drawn on number of ionizations be tried before an effective one was found, and the search
required, size of the sensitive gene volume, etc. (10). was continued by many workers. This search was
Yet the actual processes of gene mutation and chromo- encouraged by the accumulation of data which pointed
some breakage-if indeed these are essentially different to an influence of physiological conditions, such as age
from one another-are still as mysterious as ever. (13), sex (I), starvation (15), and of the genotype itself
X-rays are destructive and nondiscriminating. I t is (5, 11, 16) on mutation rate. If, thus, chemical condi-
conceivable that less severe methods of producing muta- tions created by the organism itself are capable of
tions might make possible a closer insight into the proc- influencing the process of mutation, it did not seem
esses concerned. An advance in this direction was made beyond hope that chemical substances introduced from
when it was found that ultraviolet light, too, is capable of outside might have similar. effects.
producing mutations. The restriction of the effective The choice of mustard gas for trials of this kind was
wave length to a comparatively narrow range which suggested by observations pointing to its interference
includes the absorption bands of nucleic acid and certain with cell division. Mustard gas burns, like X-rays
protein components supported the hypothesis, put bums, heal only with difficulty, and even after they
forward by radiation geneticists, that the first step appear to have healed they have a tendency to break
toward the production of a mutation consists in the .'down again. In addition, it was found, that vaginal
absorption of an energy quantum by some constituent of epithelium of an ovariectomized mouse which has been
the chromosome. But the hope that more specific exposed to a weak solution of mustard gas fails to mani-
effects, dependent on thewave length, might be produced, fest the mitotic activity which normally follows stimu-
was not realized. Certain differences between the action lation with estrogens, and that this inhibition of mitosis
of ultraviolet and that of the much shorter waves of X- lasts for several weeks after the exposure to mustard
and gamma rays are not yet understood, but may gas. I t is well known that the chromosome breaks and
eventually prove helpful in the analysis of the mutation rearrangements caused by X-radiation interfere with cell
process (18). proliferation, partly through mechanical disturbances of
Chemical substances with mutagenic properties should mitosis and partly through death of those cells which,
be particularly useful tools for attacking the problems of after distribution of the fragments and new chromosome
combinations into the daughter cells, do not receive a
sufficiently normal set of chromosomes. I t was thought
possible that mustard gas, like X-rays, may inhibit cell
division through direct action on the chromosomes.
In the autumn of 1940, experiments were started to
find out whether mustard gas is capable of producing gene
mutations and chromosome rearrangements. Drosophila
melanogastw was used as test animal. The flies were ex-
posed to mustard gas vapor, first in a closed chambe; and
later in a container through which air mixed with mustard
gas was sucked. The first results immediately gave prom-
ise of success. Both males and females became steri-
lized to a degree which depended on the dose. Sterility
was found to be due to two independent causes, both of
which are also known to be involved in the production of
X-ray sterility. First, gametogenesis is inhibited, so that
after a time no more ova and spermatozoa are available.
Second, lethality is high among zygotes from treated
eggs and, more important still, among eggs laid by un-
treated females which have been mated to treated males.
Since the spermatozoa do not lose their motility as a
result of the treatment, the most likely explanation was
that mustard gas, like X-rays, produces chromosome
breaks and rearrangements in the sperm.
In order to obtain conclusive proof that mustard gas
exerts an action on the chromosomes, genetic methods
for the detection of mutations were applied. Male
flies were mated to untreated females, and the progeny
(Fz)was examined for the occurrence of mutations.
Early mutation work, as well as some more recent work
on organisms which are genetically less thoroughly known
than Drosophila, has suffered from the impossibility of
eliminating the large personal error, for a trained worker
may spot abnormalities which may pass unnoticed by a
less experienced or less observant person. In Drosophila
genetics this obstacle has been removed by methods
which, in the main, have been designed by H. J. Muller,
and without which the quantitative analysis of genetic ra-
diation effects would have been impossible. The essential
feature of these methods is their restriction to the detec-
tion of so-called lethal mutations, i.e mutations which are
so harmful that they prevent development of the individ-
ual. Hence, lethal mutations are detected by the
absence from the progeny of a whole class of flies, and
since presence or absence are characteristics about
which any two observers are likely to agree, these
methods reduce the personal error to a minimum, while
a t the same time allowing the study of large samples
without excessive labor. Particularly useful for large-scale
tests are methods like the famous CZB test which are de-
signed to detect sex-linked lethals, i.e. lethals on the sex
chromosome, of which the male has only one, while the
female has two. A sex-linked lethal prevents the develop-
ment of a male carrying it, while it usually does not
interfere seriously with development of the female. I n the
C1B test each treated or control sex-chromosome becomes
subsequently represented, in the Fz,by a whole culture
of flies, and if a lethal has arisen on a sex-chromosome, the
corresponding culture will consist entirely of females-
a fact which is, of course, readily observed even by an
untrained person.
The result of the first C1B test with mustard gas, car-
ried out in April 1941, was spectacular beyond expecta-
tion. Whereas the rate a t which sex-linked lethals arise
spontaneously in laboratory stocks rarely approaches 1
per cent, 90 lethals were found in about 1,300 treated
sex-chromosomes. This represents a mutation rate of
over 7 per cent. Only 3 sex-linked lethals were found in
an equivalent number of untreated chromosomes, rep-
resenting a rate of 0.2 per cent. Similar results had
previously been obtained only with X-rays or other
high-energy radiation. Further. tests fully confirmed
and even exceeded the first success, up to 24 per cent
lethals being produced. Higher percentages can hardly
be expected because, concomitantly with the increase in
mutation rate, sterility becomes more and more severe.
Genetic analysis of the lethals produced in the first
C1B test indicated that some of them were due to, or
combined with, chromosome rearrangements, and these
findings were confirmed by cytological examination
carried out by Dr. Slizynski. A special test for the pro-
duction of chromosome rearrangements by mustard gas
was undertaken in December 1941. The method was
designed to spot translocations, i.e rearrangements
through which two chromosomes have exchanged por-
tions with one another. Spontaneous translocations are
so exceedingly rare that the use of controls was not con-
sidered necessary. The result left no doubt about
the capacity of mustard gas to produce chromosome
rearrangements; 7 translocations were found in 816
treated nuclei. A report on these results was sent to
the Ministry of Supply in March 1942, but, like all this
work, could not be published because of the security
ban on work with war gases. I n subsequent experiments
more translocations as well as other types of rearrange-
ments were produced. Since only Drosophila had been
used for all these studies, it was gratifying that cytological
investigations on pollen mother cells of Tradesmnlia,
carried out by Dr. Roller in 1943, fully confirmed our
finding that mustard gas can produce chromosome breaks
and rearrangements.
The similarity between the genetic effects of mustard
gas and of X-rays are so striking that only gradually did
certain differences between the two types of action come
to light. Yet special interest attaches just to these
differences, because a comparison between chemical and
physical mutagens seems a hopeful approach to the
problem of mutation. The first difference appeared in
work on translocations. I t has been shown that the
frequencies of X-ray-induced lethals, on the one hand,
and of X-ray-induced translocations, on the other, bear
a mathematical relationship to the dose administered,
the first increasing directly as the dose, the second
approximately as its 3/2th power (10). Consequently,
for a given dose of X-rays (as measured in roentgen
units) there exists a numerical relationship between the
numbers of lethals and translocations produced. Thus, a
dose of 3,000 r-units produces about 9 per cent sex-linked
lethals and about 6 per cent translocations between
chromosomes I1and I11of D. melanogaster. After mustard
gas treatment, this relationship is shifted very markedly
in favor of sex-linked lethals. Instead of the expected
6 per cent, only 0.5 per cent translocations between chro-
mosomes I1 and I11 were produced in an experiment in
which the rate of sex-linked lethals was 9 per cent, and a
siinilar relative shortage of translocations was observed
in subsequent tests. At first sight, these observations
seem to indicate that mustard gas is less efficient than
X-rays in breaking the chromosome thread. However,
it is well to be cautious in drawing this conclusion. It
has to be kept in mind that with the methods used we
could not detect the primary breaks, but only a pro-
portion of the subsequently formed rearrangements. I t
is conceivable that chemical treatment interferes with
the process of rejoining of broken ends in such a way that
a given number of breaks results in fewer observable
rearrangements than would be formed by the same
number of X-ray breaks. Special tests are required to
decide this point.
On the other hand, mustard gas does not seem to be
less efficient than X-rays in the production of very
small, so-called "minute" rearrangements. Slizynski and
Slizynska (17),in a cytological of sex-linked lethals pro-
duced by various agencies, have found that in about 20
per cent of cases the genetic change underlying the pro-
duction of a lethal is a minute deficiency in the chro-
mosome, and this frequency appears to be the same after
X-rays, after ultraviolet radiation, and after mustard
gas treatment. These findings emphasize the similarity,
often pointed out by geneticists, between true gene
mutations and minute chromosome rearrangements, and
they do not contradict the possibility that so-called gene
mutations may be nothing more than chromosome
rearrangements of so minute a size that they elude
detection by cytological methods. I t will be of great
interest to determine whether small deficiencies form an
equally high proportion of lethals which have been pro-
duced by less potent chemical substances than mustard
gas.
A second difference between the actions of X-rays and
of mustard gas came to light in the course of a study of
visible mutations after chemical treatment. In one
respect this study was disappointing since the mutations
observed were of the same types as those found after
X-ray treatment, and there was no indication of any
specific effects of the gas on individual genes. Mustard
gas seems to act as indiscriminately as X-radiation.
There is, however, a differenke between these two
agencies which involves not the types of mutation which
they produce, but the way in which the mutations
become manifest in the offspring of the treated flies.
After X-ray treatment of males most of the mutated
offspring show the induced abnormality (such as yellow
body color instead of the normal gray) over the whole
surface of their body. Only a small proportion (less
than 15 per cent) of the mutated individuals are mosaics,
i.e. show the abnormality in a part of their body, the
remainder being normal. In the progeny of mustard-
gas-treated males, on the other hand, mosaics form a
high proportion (usually between 30 and 50 per cent) of
all mutated individuals ( 2 ) . Moreover, whereas the
gonads of X-ray mosaics rarely contain both normal and
mutated cells, those of mustard gas mosaics quite fre-
quently appear to do so. A special study has been made
of such "gonadic mosaicism" with respect to sex-linked
lethal mutations (3). A female, daughter of a treated
male, whose ovaries contain a patch of tissue in which
the cells carry a sex-linked lethal, will have fewer sons
than a normal female, the shortage of sons depending on
the relative sizes of the normal and mutated portions of
the ovary. A similar depression of the sex ratio occurs
also in the progeny of females who carry, evenly distrib-
uted through all cells of their ovaries, a sex-linked
"semilethal" mutation, i.e. a mutation which weakens the
males so that only a proportion of them are capable of
completing development. Analysis of females giving
a low sex ratio has shown that among daughters of ir-
radiated males this abnormality is almost always due to
a semilethal mutation affecting the whole of the ovary.
On the other hand, in 9out of 20 daughters of mustard-gas-
treated males, the shortage of sons was due to the
presence in their otherwise normal ovaries of a mosaic
patch carrying a sex-linked lethal. Finally, mention
should also be made of a striking case of mosaicism in
which a son of a mustard-gas-treated male was, both in
the gonads and in the soma, a mosiac for two different
mutations of the same gene, although it must be assumed
that in the treated spermatozoon each treated gene was
present only once.
An explanation which seems particularly satisfactory
in accounting for all these observations is that the gene
affectedby treatment does not always mutate at once, but
may acquire a tendency to mutate which remains latent
until a later cell division. Support for this hypothesis
was obtained when it was found in several cases that the
offspring of gonadic mosaics for a mutation again were
gonadic mosaics for the same mutation. In these cases
an induced specific instability seems to have been trans-
mitted from one generation to the next before giving rise
to a stable change. No parallel observations have been
reported in literature on radiation genetics; but it seems
worth noting that so-called "unstable" genes, i.e. genes
which tend to mutate repeatedly in the same direction,
have been found several times in untreated material (6).
 The difference between the mutations produced by
short-wave radiation and chemical reaction may be
tentatively ascribed to the different amounts of energy
involved in the two types of reaction. I n short-wave
radiation the energy made available is usually suffi-
ciently large to produce a catastrophic alteration in the
structure of the gene, by transforming it from one stable
configuration to another. On the other hand, reaction of
the gene with a chemical substance, because of the smaller
amount of energy involved, may produce a less drastic
effect, by transforming it to an intermediate metastable
configuration. Such a configuration will, of course, tend
to undergo "spontaneous"a1teration to another and more
stable configuration, i.e. an "unstable" gene is produced.
After the first positive results with mustard gas had
been obtained, the search for chemical mutagens was
extended to substances which either in their chemical
structure or in their pharmacological action are related
to mustard gas. Mustard gas is a fixative of protoplasm
with unusual power of penetration. I t was soon seen
that these two properties by themselves are not suffi-
cient to make a substance mutagenic, for neither osmic
acid nor picric acid affected the mutation rate in tests in
which the majority of the treated individuals were killed.
Neither can it be said that every potent vesicant is a
mutagen, for lewisite gave entirely negative results in two
ClB tests. So far only three substances have been found
which give genetic effects similar to those of mustard
gas. These all belong to the class of nitrogen- or sulphur-
mustards. Their chemical formulas are: (1) O(CH2.
CH2.S. CH2. CHzC1)2, (2) N(CH2. CHzCl),, and (3) CHI.
N(CH2. CH2Cl)z.
As the chemical mutagens presumably attack the genic
material directly, it was noted with interest that these
active compounds all contain an unsaturated atom
(:S or i N) which might combine with materials composing
the gene, and that this activity would be enhanced by
the type of side chains present in the vesicant musbrds
(e.g. C1CH2.CH2-). With this type of structure is asso-
ciated the tendency to intramolecular cyclization to
form onium compounds (7), characteristic of the active
tN and :S vesicants. Assuming that the mutagenic
action is due to the reaction on the unsaturated atom of
the uncyclized compound with the gene, it was thought
likely that replacement of the :S and i N by :0-wouldnot
produce active materials, as the addition compounds of
divalent oxygen are not stable in aqueous solution (and
"0 mustard" is not a vesicant). Stable addition com-
pounds are, however, formed by compounds having the
:CO group, and though the corresponding "CO
mustards" were not tried, the tear gases, chloracetone
and dichloracetone (compounds with C1CH2-side chains
instead of the C1CH2.CH2- of mustard gas) were
examined. The activity of these compounds was
extremely weak; in fact, the results were not clearly
positive, although they suggest the possibility that both
these compounds are slightly mutagenic. By analogy
with the vesicant action of the :S compounds, the
chlorethylacetones might be more active. I t has not
yet been possible to test these substances.
Other groups which might replace the :S and i N of the
mustards and retain the additive ability of the molecule
include :As, although compounds with i As are rather
unstable under physiological conditions and are mostly
toxic. The corresponding i P compounds are apparently
out of the question in this connection, as they are spon-
taneously inflammable in air. The :SOzcompounds may
similarly be expected to show some activity, but not the
:SO ones. This parallels their vesicant action. There
is also the possibility of using nitrile, isonitrile, and the
corresponding thiocyanates and isothiocyanates for the
coordinating group, and attention was therefore directed
to ally1 isothiocyanate. A weak but definite mutagenic
activity could be demonstrated for this compound (4).
On the other hand, this type of chemical structure does
not seem a necessary prerequisite for a mutagenic sub-
stance. This is shown by the fact that Hadorn and
Niggli (8) have obtained considerable numbers of
mutations by exposing explanted ovaries of Drosophila
to weak solutions of ~henol.
Ally1 isothiocyanate, or mustard oil, occurs naturally
in plants of the genus Brassica. We do not know
whether it acts as a mutagen in these plants, but it is
interesting to speculate how far naturally occurring
mutagens may be responsible for spontaneous mutability.
I t has been shown by Muller and Mott-Smith (14) that
cosmic radiation and natural radioactivity are quanti-
tatively insufficient to account for the observed rates of
spontaneous mutation. Timofkeff-Ressovsky, Zimmer,
and Delbriick (20) have suggested that random tempera-
ture oscillations inside the nucleus may occasionally
overstep the energy threshold required for the production
of a mutation. In the light of the results reported here
it seems possible, however, that a certain proportion of
natural mutations may be due to the action of mutagenic
substances within the organism, and whose production
may itself be the consequence of gene action. Such an
assumption finds support in the known cases in which
definite genes influence mutability of the rest of the genes
or even of a specific gene (2, 11, 17). Search for natural
mutagens is therefore of high interest, but may well turn
out to be exceedingly difficult. By its very nature a
natural mutagen can have no drastic effect in the species
in which it occurs; otherwise, the species could not
survive. Moreover, a physiological system which in-
cludes mutagens whose production is controlled by genes
will, in the course of its evolution, have attained a finely
attuned equilibrium between the strength of the effective
substances and the sensitivity of the gene or genes on
which they act. Removed from its normal genotypical
environment, a natural mutagen may produce quite
different effects or none a t all. Therefore, results
obtained with one organism may not be transferable to
another, in contrast to results gained with such drastic
agencies as X-rays and, presumably, mustard gas.
Nevertheless, it is tempting to consider the possibility
that one of the means by which evolution adapts mut-
ability to environmental 'requirements is the achieve-
ment of a balance between the production of mutagens
and sensitivity to them.
References
1. AUERBACH, C. I . Genet., 1941,41,255-265.
2 . AUERBACH, C. Proc. roy SOC.Edinb., 1916, 62, 211-221.
3. AUERBACH, C. PYOC.roy. SOC.Edinh., in press.
4. AUERBACH, C., and ROBSON,J. M. Nature, Lond., 1944, 154, 81-82.
5. D E ~ X E M.
C , Genetics, 1937,22,469-478.
Henry Helm Clayton
1861-1946
With the passing of Henry Helm Clayton, on October
26, 1946, there ended a life of exceptional activity and
eminence in meteorology, public service, and business.
Clayton was born on March 12, 1861, a t Murfreesboro,
Tennessee. Because of delicate health, early education
was acquired in private schools and by study a t home.
I t was during this period that his interest in meteor-
ology developed.
Studies of local storms, beginning in 1878, were
followed in 1882 by his first activity, aid in the organ-
ization of the newly formed Tennessee Weather Service,
including analyses of reports and a gift of 30 rain gauges.
In 1884-85 he was assistant a t the Observatory of the
University of Michigan and associate editor of the
American Meteorological Journal. In February 1886, after
three months a t Harvard College Observatory, he joined
the staff of Blue HillMeteorologicalObservatory, founded
by Abbott Lawrence Rotch in the preceding year, where
he remained as assistant and meteorologist until 1909.
During this period many important advances were
initiated by him with the enthusiastic approval of Direc-
tor Rotch. At first there was no assistant, and a t various
times, at his own expense, Clayton employed others to
share his rapidly expanding program of research.
Clayton's studies of clouds, begun in 1886, yielded the
first definite information concerning the circulation of
the atmosphere over America and established the Clayton-
Egnell law of the increase of velocity with height.
The detailed observations during the period 1886-90
are the only hourly data of changes in form, height, and
movement of clouds in the Western Hemisphere, and,
with Clayton's discussion, undoubtedly stimulated the
6. DEmREC, Me S~mfios. Want. Bio1.P 1941,9,145-150.
. , PHILIPS,F. S. Science, 1946,103, 409-4'15.
7. G I ~ A N , A and
8. HADORN, E., and NIDDLI,H. Nature, LO&.. 1946,157.162-163.
9. LAW, L. W. Proc. nat. cad. Sci. wash., 1938, ?4;5461550.
10. LEA,D . E. Actions of radiations on living cells. Cambridge, Engl.: Cam.
bridge Univ. Press, 1946.
11. MAYPELL,K. PIOC.nat. Acad. Sci. Wash., 1943,29, 137-143.
12. MULLER,H. J. Verh. V . int. Kongr. Vererbungsw., 1927, 234-260; 2.
indukt. Abstamm. Vererb Lehre (Suppl. I ) , 1928.
13. MULLER,H. J: Genetics,1946,31,225.
14. MULLER,H. J., and MOTT-SXITH, L. M. PIOC.mat. Acad. Sci. TVash.,
1930,16, 277-285,
15. OLENOV, J. M. Amer. Nat., 1941,75,580-595.
16. RHOADES, M.M. Symfios. quanl. Biol., 1941, 9, 138-144.
17. SLIZYNSKI, B. M., and SLIZYNSKA, H. PIOC.TOY. Sot. Edinb., in press.
18. STADLER,L. J. Symfios. quant. Biol., 1941,9,168-178.
19. TIMOF~EFP-RESSOVSKY, N. W. Biol. Rev., 1934,9, 411-457.
20. TIMOF~RPF-RESSOVSKY,N. W., ZI=R,K. G.,and D ~ ~ s n i t c k , M .
Nochr.
Ges. TViss. Gfittingen (Math.-Phys. Kl., Biol.) 1935, 1, 234-241.
organization of the International Cloud Committee and
the international series of measurements of heights and
velocities of 1896. The colored pictures for the Atlas of
clods, issued in 1897 by the U. S. Hydrographic Office,
were painted under Clayton's supervision. He was also
consultant in the preparation of the first Iflter9tational
atlas (1905).
Clayton's invitation to William A. Eddy to try his
meteorological kites a t Blue Hill led to the first use of
kites to lift recording instruments, August 4,1894, and to
the adoption throughout the world of this method of
sounding the atmosphere. Important results of Clayton's
analyses of the accurate data obtained included the
discovery of persistent, sharply defined stratifications in
the lower atmosphere, and previously unsuspected, vari-
able effects of mountains upon the surrounding atmos-
phere. Interest in this new method of research, later to be
named "aerology," is indicated by his generosity in allow-
ing free use in aerology, without royalty, of the form of
Hargrave kite patented by him, which came into use a t
all aeronautical laboratories. Always envisaging improved
techniques and advances in aerology, he encouraged Blue
Hill to develop the first radio sonde, in 1935.
He described the solar eclipse as "a kind of laboratory
experiment in which are eliminated practically all in-
fluences upon the atmosphere except that of a fall of
temperature," and original studies led to his suggestion
of the eclipse cyclone caused by the cooling of the air
by the shadow.
Clayton's great interest in forecasting was first in-
dicated by his paper, "A lately-discovered meteorological
cycle," published in the Amer6cam Me~eorologicalJournal
in August 1884. Later, a t Blue Hill, his proof that fore-
casts made locally are better than those issued a t a